News sites including The Guardian, MailOnline and The Times have all suggested this week that the risk of severe coronavirus illness for black and minority ethnic (BME) patients could be three times higher than that facing white patients. Though all quoted an expert who warned that more research is needed before drawing definitive conclusions.
The stories are based on a report by the Intensive Care National Audit and Research Centre (ICNARC).
On closer inspection, the data does not show that ethnic minority patients overall are three times more likely to need critical care for coronavirus than their white neighbours.
It does suggest black patients in particular are more likely to be in intensive care than local demographics would predict – though there may be other factors in play.
The ICNARC report looks at 2,249 intensive care patients with coronavirus in the UK. Some 65 per cent were white, and the remaining 35 per cent were from ethnic minorities.
Journalists have compared that 35 per cent figure with the proportion of the UK population who said they belonged to an ethnic minority in the 2011 census (13 per cent).
But this is an unhelpful comparison.
One thing we know about coronavirus is that it is hitting urban areas – especially London – first and hardest. And these big cities are more likely to have large BME populations.
To adjust for this, the ICNARC researchers compared the ethnicity of Covid-19 patients with the make-up of the local authority wards those patients live in.
According to the ICNARC analysis, 65 per cent of Covid-19 patients receiving critical care are white, compared to 74 per cent of residents in their local area.
The proportion of Asian patients in ICU is, this data suggests, almost exactly in line with the demography of the area they live – at around 13 per cent in each.
However, the same report seems to show that black patients are significantly more likely to be critically ill with Covid-19 than local demography would suggest. Some 14 per cent of ICU patients are black, compared to 7 per cent of the local authority wards in which those patients live.
The limits of the data
The news stories and the ICNARC analysis all use the 2011 census to make their comparisons between the ethnicity of intensive care patients and the local population.
Those statistics are the latest available census data, but they are now nearly ten years old.
Our analysis of more recent ONS figures shows that between 2012 and 2018, the proportion of London residents who were black rose from 10.9 per cent to 11.9 per cent.
In the West Midlands, which is also experiencing a significant coronavirus outbreak, that figure rose from 2.9 per cent to 4.9 per cent over the same period.
The proportion of residents who are white also fell in those two areas over the same time (down one percentage point in London and two points in the West Midlands).
These don’t sound like huge changes, but at the level of individual council wards, can make a big difference.
Unfortunately we can’t check the ICNARC findings against the newer ONS data because we don’t have information on individual ICU patients to compare it to.
But it is possible that at least some of the apparent over-representation of black patients in critical care could be explained by wider demographic shifts in the places these individuals live.
News articles this week have suggested ethnic minority patients with coronavirus could be three times more likely to end up in intensive care.
But nothing in the research these stories are based on supports the idea that BME patients face a risk three times higher than their white neighbours – once we account for the areas patients live.
Compared to the make-up of council wards they live in, white coronavirus patients were slightly under-represented in critical care, the proportion of Asian patients was in line with what local demography predicts. Black coronavirus patients were significantly over-represented in ICU, according to these figures.
The researchers compared patient profiles with ethnicity data collected in the 2011 census. It’s possible some or all of the disparities are explained by local demographic changes in the intervening years – though for now, we cannot say definitively either way.
It’s true that BME people are more likely to live in urban areas and therefore more likely to have already been exposed to Covid-19. These are the places that have been hit first and hardest so far. As the news articles we’ve looked at made clear, more research in this area is needed.